A partial volume correction-based approach for detection of receptor/transporter distribution in the human brain: Initial application to [11C]raclopride PET scans

Background and aims: Voxel-wise identification of receptor/transporter distributions in the brain with PET may be confounded for regions of relatively low densities or when only suboptimal ligands are available for the site. We tested whether a new partial volume correction (PVC)-based approach identifies regions of relatively low densities of dopamine D2/D3 receptors with [11C]raclopride, a relatively low affinity ligand, compared to published Results: using higher affinity D2/D3 radioligands or autoradiography. Methods: Twenty-one normal subjects (age: 31 ± 8) underwent one 90-min dynamic [11C]raclopride scan. Binding potential (BP) maps were constructed by the bolus-plus-infusion transformation (BPIT; Kuwabara et al., 2002). For PVC, tissue segments were identified on MRI, transferred to PET space, and smoothed by PET camera specific Gaussian kernels to reproduce the smearing effect. The resulting geometric transformation matrix (M, Rousset et al., 1998), is an n by m matrix where n and m are numbers of voxels to perform PVC and tissue segments, respectively. PV-corrected BP values of tissue segments (T, an m by 1 matrix) were obtained by left matrix division (M\T) using Gaussian elimination. The PVC-predicted BP map was constructed by inserting the M*T product to the PVC voxels and -1 to the remaining voxels. Manually defined VOIs for putamen, caudate nucleus, and cerebellum, and automatically defined (using SPM2) white matter, gray matter, and cerebrospinal space segments (m=8) were used for the initial iteration. PVC-predicted BP maps represent weighted mean BP values of the segments at the voxel level assuming a uniform BP value for each segment. Comparison of observed BP maps to PVC-predicted BP maps, using SPM2, allows exploration of regions of significant deviation from the homogeneity assumption. New segments that correspond to regions of significant deviation were included for the second iteration as described below. Results: The first iteration of SPM analysis identified clusters in posterior putamen (peak t=33.95 at [-28,-10,6]), ventral striatum (t=-11.64 at [20,8,0]), thalamus (t=25.23 at [4,-14,8]), globus pallidus externa (t=19.66 at [20,-8,8]), internal capsule between putamen and caudate nucleus (t=13.94 at [-22,14,14]), and rectus gyrus (t=13.32 at [20,22,-12]) bilaterally at a p<0.05 level, corrected for multiple comparisons and setting the minimal cluster volume to 1 ml. Inclusion of above clusters as separate tissue segments in a second iteration of PVC (m=20) revealed no clusters in SPM analysis at the same criteria. The identified regions generally agreed with those reported using high affinity ligand [18F]fallypride (e.g., Mukherjee et al., 2002) except for amygdala. The globus pallidus externa and internal capsule clusters, which were not reported with [18F]fallypride, are supported by in vivo autoradiography (e.g., Tupla et al., 2003). Conclusions: The proposed approach successfully identified extra striatal region with moderate dopamine D2/D3 receptor density with [11C]raclopride PET that agreed with published Results: using higher affinity ligands and in vivo materials. Although the approach appears promising, further validation will be required to apply this method to receptor and transporter studies with diffuse distributions, as well as to identify their changes in clinical populations.